I am developing an application that is responsible of moving and managing robots over an UDP connection.
The application needs to:
Read joystick/user input using SDL.
Generate and send a control packet to the robot every 20 milliseconds (UDP)
Receive and decode response packets from the robot (~20 msecs). This was implemented with the signal/slot mechanism and does not require a timer.
Receive and process robot messages for debugging reasons. This is not time-regulated.
Update the UI regularly to keep the user notified about the status of the robot (e.g. battery voltage). For most cases, I have also used Qt's signal/slot mechanism.
Use a watchdog that disables the robot if no response is received after 1 second. The watchdog is reset when the application receives a robot packet (~20 msecs)
For the moment, I have implemented all of the above. However, the application fails to send the packets regularly when the watchdog is activated or when two or more QTimer objects are used. The application would generally work, but I would not consider it "production ready". I have tried to use the precision flags of the timers (Qt::Precise, Qt::Coarse and Qt::VeryCoarse), but I still experienced problems.
Notes:
The code is generally well organized, there are no "god objects" in the code base (most source files are less than 150 lines long and only create the necessary dependencies).
Most of the times, I use QTimer::singleShot() (e.g. I will only send the next packet once the current packet has been sent).
Where we use timers:
To read joystick input (~50 msecs, precise timer)
To send robot packets (~20 msecs, precise timer)
To update some aspects of the UI (~500 msecs, coarse timer)
To update the elapsed time since the robot was enabled (~100 msecs, precise timer)
To implement a watchdog (put the application and robot in safe state if 1000 msecs have passed without a robot response)
Note: the watchdog is feed when we receive a response packet from the robot (~20 msecs)
Do you have any recommendations for using QTimer objects with performance-critical code (any idea is welcome). Note that I have also tried to use different threads, but it has caused me more problems, since the application would not be in "sync", thus failing to effectively control the robots that we have tested.
Actually, I seem to have underestimated Qt's timer and event loop performance. On my system I get on average around 20k nanoseconds for an event loop cycle plus the overhead from scheduling a queued function call, and a timer with interval 1 millisecond is rarely late, most of the timeouts are a few thousand nanoseconds short of a millisecond. But it is a high end system, on embedded hardware it may be a lot worse.
You should take the time and profile your target system and Qt build to determine whether it can indeed run snappy enough, and based on those measurements, adjust your timings to compensate for the system delays to get your events scheduled more on time.
You should definitely keep the timer thread as free as possible, because if you block it by IO or extensive computation, your timer will not be accurate. Use a dedicated thread to schedule work and extra worker threads to do the actual work. You may also try playing with thread priorities a bit.
Worst case scenario, look for 3rd party high performance event loop implementations or create your own and potentially, also a faster signaling mechanism as ell. As I already mentioned in the comments, Qt's inter-thread queued signals are very slow, at least compared to something like indirect function calls.
Last but not least, if you want to do task X every N units of time, it will only be only possible if task X takes N units of time or less on your system. You need to make this consideration for each task, and for all tasks running concurrently. And in order to get accurate scheduling, you should measure how long did task X took, and if less than its frequency, schedule the next execution in the time remaining, otherwise execute immediately.
Related
New description of the problem:
I currently run our new data acquisition software in a test environment. The software has two main threads. One contains a fast loop which communicates with the hardware and pushes the data into a dual buffer. Every few seconds, this loop freezes for 200 ms. I did several tests but none of them let me figure out what the software is waiting for. Since the software is rather complex and the test environment could interfere too with the software, I need a tool/technique to test what the recorder thread is waiting for while it is blocked for 200 ms. What tool would be useful to achieve this?
Original question:
In our data acquisition software, we have two threads that provide the main functionality. One thread is responsible for collecting the data from the different sensors and a second thread saves the data to disc in big blocks. The data is collected in a double buffer. It typically contains 100000 bytes per item and collects up to 300 items per second. One buffer is used to write to in the data collection thread and one buffer is used to read the data and save it to disc in the second thread. If all the data has been read, the buffers are switched. The switch of the buffers seems to be a major performance problem. Each time the buffer switches, the data collection thread blocks for about 200 ms, which is far too long. However, it happens once in a while, that the switching is much faster, taking nearly no time at all. (Test PC: Windows 7 64 bit, i5-4570 CPU #3.2 GHz (4 cores), 16 GB DDR3 (800 MHz)).
My guess is, that the performance problem is linked to the data being exchanged between cores. Only if the threads run on the same core by chance, the exchange would be much faster. I thought about setting the thread affinity mask in a way to force both threads to run on the same core, but this also means, that I lose real parallelism. Another idea was to let the buffers collect more data before switching, but this dramatically reduces the update frequency of the data display, since it has to wait for the buffer to switch before it can access the new data.
My question is: Is there a technique to move data from one thread to another which does not disturb the collection thread?
Edit: The double buffer is implemented as two std::vectors which are used as ring buffers. A bool (int) variable is used to tell which buffer is the active write buffer. Each time the double buffer is accessed, the bool value is checked to know which vector should be used. Switching the buffers in the double buffer just means toggling this bool value. Of course during the toggling all reading and writing is blocked by a mutex. I don't think that this mutex could possibly be blocking for 200 ms. By the way, the 200 ms are very reproducible for each switch event.
Locking and releasing a mutex just to switch one bool variable will not take 200ms.
Main problem is probably that two threads are blocking each other in some way.
This kind of blocking is called lock contention. Basically this occurs whenever one process or thread attempts to acquire a lock held by another process or thread. Instead parallelism you have two thread waiting for each other to finish their part of work, having similar effect as in single threaded approach.
For further reading I recommend this article for a read, which describes lock contention with more detailed level.
Since you are running on windows maybe you use visual studio? if yes I would resort to VS profiler which is quite good (IMHO) in such cases, once you don't need to check data/instruction caches (then the Intel's vTune is a natural choice). From my experience VS is good enough to catch contention problems as well as CPU bottlenecks. you can run it directly from VS or as standalone tool. you don't need the VS installed on your test machine you can just copy the tool and run it locally.
VSPerfCmd.exe /start:SAMPLE /attach:12345 /output:samples - attach to process 12345 and gather CPU sampling info
VSPerfCmd.exe /detach:12345 - detach from process
VSPerfCmd.exe /shutdown - shutdown the profiler, the samples.vsp is written (see first line)
then you can open the file and inspect it in visual studio. if you don't see anything making your CPU busy switch to contention profiling - just change the "start" argument from "SAMPLE" to "CONCURRENCY"
The tool is located under %YourVSInstallDir%\Team Tools\Performance Tools\, AFAIR it is available from VS2010
Good luck
After discussing the problem in the chat, it turned out that the Windows Performance Analyser is a suitable tool to use. The software is part of the Windows SDK and can be opened using the command wprui in a command window. (Alois Kraus posted this useful link: http://geekswithblogs.net/akraus1/archive/2014/04/30/156156.aspx in the chat). The following steps revealed what the software had been waiting on:
Record information with the WPR using the default settings and load the saved file in the WPA.
Identify the relevant thread. In this case, the recording thread and the saving thread obviously had the highest CPU load. The saving thread could be easily identified. Since it saves data to disc, it is the one that with file access. (Look at Memory->Hard Faults)
Check out Computation->CPU usage (Precise) and select Utilization by Process, Thread. Select the process you are analysing. Best display the columns in the order: NewProcess, ReadyingProcess, ReadyingThreadId, NewThreadID, [yellow bar], Ready (µs) sum, Wait(µs) sum, Count...
Under ReadyingProcess, I looked for the process with the largest Wait (µs) since I expected this one to be responsible for the delays.
Under ReadyingThreadID I checked each line referring to the thread with the delays in the NewThreadId column. After a short search, I found a thread that showed frequent Waits of about 100 ms, which always showed up as a pair. In the column ReadyingThreadID, I was able to read the id of the thread the recording loop was waiting for.
According to its CPU usage, this thread did basically nothing. In our special case, this led me to the assumption that the serial port io command could cause this wait. After deactivating them, the delay was gone. The important discovery was that the 200 ms delay was in fact composed of two 100 ms delays.
Further analysis showed that the fetch data command via the virtual serial port pair gets sometimes lost. This might be linked to very high CPU load in the data saving and compression loop. If the fetch command gets lost, no data is received and the first as well as the second attempt to receive the data timed out with their 100 ms timeout time.
Windows Embedded Compact 7.
Is there a way to test interrupt latency time from user space?
Are there any tools provided as part of platform builder?
I also saw a program called Intrtime.exe - but no examples on how to use it.
How does one test the interrupt latency time?
Reference for Intrtime.exe but how do I implement it?
http://www.ece.ufrgs.br/~cpereira/temporeal_pos/www/WindowsCE2RT.htm
EDIT
Also found:
ILTiming.exe Real-Time Measurement Tool (Compact 2013)
http://msdn.microsoft.com/en-us/library/ee483144.aspx
This really is a test that requires hardware, and there are a couple "latencies" you might measure. Once is the time from the interrupt signal to when the driver ISR reacts and the second is from when the interrupt occurs to when an IST reacts.
I did this back in the CE 3.0/CE 4.0 days by attaching a signal generator to an interruptable input an then having an ISR pulse a second input and an IST pulse a third input when they received the interrupt. I hooked a scope up to the input and outputs and used it to measure time between the input signal and output signals to get not just latency, but also jitter. You could easily add a 4th line for CE 7 so you could check an IST in user space and an IST in kernel space. I'd definitely be interested to see the results.
I don't think you can effectively measure this with software running on the platform, as you get into the problem of the code trying to do the measurement affecting the results. You're also talking time way, way below the system tick resolution so the scheduler is going to be problematic as well. CeLog might be able to get you an idea on these times, but getting it set up and running is probably more work than just hooking up a scope.
What is usually meant by interrupt latency is the time between an interrupt source asserting the interrupt line and a thread (sometimes in user-space) being scheduled and then executing as a result.
Unless your CPU has some accurate way of time-stamping interrupt events as they arrive at the CPU (rather than when an ISR runs), the only truly accurate measurement is one done externally - by measuring the time between a the interrupt line being asserted and some observable signal that the thread responding to the interrupt can control. A DSO or logic analyser is usually used for this purpose.
Software techniques usually rely on storing an accurate time-stamp at the earliest opportunity in an ISR. If you're certain the time between interrupt line becoming asserted and the ISR running is negligible, this might be valid. If, on the other hand, disabling of interrupts is being used to control concurrency, or interrupts are nested, you probably want to be measuring this as well.
I have a data acquisition application running on Windows 7, using VC2010 in C++. One thread is a heartbeat which sends out a change every .2 seconds to keep-alive some hardware which has a timeout of about .9 seconds. Typically the heartbeat call takes 10-20ms and the thread spends the rest of the time sleeping.
Occasionally however there will be a delay of 1-2 seconds and the hardware will shut down momentarily. The heartbeat thread is running at THREAD_PRIORITY_TIME_CRITICAL which is 15 for a normal priority process. My other threads are running at normal priority, although I use a DLL to control some other hardware and have noticed with Process Explorer that it starts several threads running at level 15.
I can't track down the source of the slow down but other theads in my application are seeing the same kind of delays when this happens. I have made several optimizations to the heartbeat code even though it is quite simple, but the occasional failures are still happening. Now I wonder if I can increase the priority of this thread beyond 15 without specifying REALTIME_PRIORITY_CLASS for the entire process. If not, are there any downsides I should be aware of to using REALTIME_PRIORITY_CLASS? (Other than this heartbeat thread, the rest of the application doesn't have real-time timing needs.)
(Or does anyone have any ideas about how to track down these slowdowns...not sure if the source could be in my app or somewhere else on the system).
Update: So I hadn't actually tried passing 31 into my AfxBeginThread call and turns out it ignores that value and sets the thread to normal priority instead of the 15 that I get with THREAD_PRIORITY_TIME_CRITICAL.
Update: Turns out running the Disk Defragmenter is a good way to cause lots of thread delays. Even running the process at REALTIME_PRIORITY_CLASS and the heartbeat thread at THREAD_PRIORITY_TIME_CRITICAL (level 31) doesn't seem to help. Next thing to try is calling AvSetMmThreadCharacteristics("Pro Audio")
Update: Scheduling heartbeat thread as "Pro Audio" does work to increase the thread's priority beyond 15 (Base=1, Dynamic=24) but it doesn't seem to make any real difference when defrag is running. I've been able to correlate many of the slowdowns with the disk defragmenter so turned off the weekly scan. Still can't explain some delays so we're going to increase to a 5-10 second watchdog timeout.
Even if you could, increasing the priority will not help. The highest priority runnable thread gets the processor at all times.
Most likely there is some extended interrupt processing occurring while interrupts are disabled. Interrupts effectively work at a higher priority than any thread.
It could be video, network, disk, serial, USB, etc., etc. It will take some insight to selectively disable or use an alternate driver to see if the problem system hesitation is affected. Once you find that, then figuring out a way to prevent it might range from trivial to impossible depending on what it is.
Without more knowledge about the system, it is hard to say. Have you tried running it on a different PC?
Officially you can't use REALTIME threads in a process which does not have the REALTIME_PRIORITY_CLASS.
Unoficially you could play with the undocumented NtSetInformationThread
see:
http://undocumented.ntinternals.net/UserMode/Undocumented%20Functions/NT%20Objects/Thread/NtSetInformationThread.html
But since I have not tried it, I don't have any more info about this.
On the other hand, as it was said before, you can never be sure that the OS will not take its time when your thread's quantum will expire. Certain poorly written drivers are often the cause of such latency.
Otherwise there is a software which can tell you if you have misbehaving kernel parts:
http://www.thesycon.de/deu/latency_check.shtml
I would try using CreateWaitableTimer() & SetWaitableTimer() and see if they are subject to the same preemption problems.
The title really says it all.
The and ... means also include pselect and ppoll..
The server project I'm working on basically structured with multiple threads. Each
thread handles one or more sessions. All the threads are identical. The protocol
takes care of which thread will host the session.
I'm using an inhouse socket class that wraps things up. The point of interest is a checkread call which calls either poll (linux) or select (windows).
In summary each thread currently calls poll on a single socket. From what I can tell, using epoll would only be of benefit if this thread was looking at multiple sockets such as what you'd get in say an HTTP server. That's not what I'm doing in my case. And the class only handles a single socket at a time.
There is some brief discussion about edge and level triggering in the man pages for epoll. I'm not really sure what it means. In the socket class I see an optimization in the windows part of the code that shortcuts the select call with an ioctlsocket & FIONREAD to check if there is any data. Wondering if that would return > 0 even if a complete UDP packet hadn't arrived at the time of the call. Is this what edge triggering is in epoll?
In some rudimentary testing, I'm also seeing no noticeable difference between using select and poll.
I can see that using ppoll might be of benefit though due to greater precision in the timeout. Any thoughts?
And yes, I am trying to optimize throughput for a session that is receiving lots of data. The server is more Network & Disk bound than CPU.
The main difference between epoll vs select or poll is that epoll scales a lot better when run in a single thread. I don't know how this would compare to using a multithreaded server using select or poll.
Look at this http://monkey.org/~provos/libevent/libevent-benchmark2.jpg
The reason for this(as far as I can tell) is that when you are using select or poll you must loop through all the connected sockets to determine which ones have data to be read. When you are using epoll, it keeps a seperate array which contains references only to sockets which have data to be read. This saves you lots of loop cycles, and the difference becomes more and more noticeable the more sockets that are connected.
Another thing to look into if performance ever becomes a major issue is io completion ports(windows only) and kqueue(FreeBSD only). It's also important to remember that epoll is linux only. In most cases select or poll will work just fine.
In the case of a single file descriptor, select and poll are more efficient than epoll due to being much simpler. (epoll has some overhead which doesn't make itself useful with only a single socket)
According to the link: http://www.intelliproject.net/articles/showArticle/index/io_multiplexing.
If you use only one descriptor:
select: 201 micro seconds.
poll: 159 micro seconds.
epoll: 176 micro seconds.
Seems poll will be a better solution in such situation.
If you have only a single socket, what's the point of polling in the first place? Wouldn't the best performance then be by just using blocking read/write?
Wrt. the performance, with only a single file descriptor I don't think there is much, if any, difference between the various approaches. If you really care, I suppose you could measure, but I find it difficult that this would particularly matter for the overall performance of your program.
Level/edge triggering. Consider you're monitoring a signal, for simplicity say some voltage in a line. Edge triggering means that something triggers when the voltage goes over or under some specific limit. Level triggering means that something is considered to be in a triggered state as long as the voltage is over/under the limit. That is, edge triggering triggers when some event happens (crossing some threshold), level triggering reflects the state of some "thing" (in this case, voltage).
To get back to network programming, and edge triggered system might be one where you get some kind of signal when a packet is received. If you don't handle the event then the signal is lost. A level triggered system, OTOH, is something like asking "is there data waiting in the buffer for me?"; if you don't handle the event and ask again, the data will still be there waiting for you.
I have a file of data Dump, in with different timestamped data available, I get the time from timestamp and sleep my c thread for that time. But the problem is that The actual time difference is 10 second and the data which I receive at the receiving end is almost 14, 15 second delay. I am using window OS. Kindly guide me.
Sorry for my week English.
The sleep function will sleep for at least as long as the time you specify, but there is no guarantee that it won't sleep for longer.If you need an accurate interval, you will need to use some other mechanism.
If I understand well:
you have a thread that send data (through network ? what is the source of data ?)
you slow down sending rythm using sleep
the received data (at the other end of network) can be delayed much more (15 s instead of 10s)
If the above describe what you are doing, your design has several flaws:
sleep is very imprecise, it will wait at least n seconds, but it may be more (especially if your system is loaded by other running apps).
networks introduce a buffering delay, you have no guarantee that your data will be send immediately on the wire (usually it is not).
the trip itself introduce some delay (latency), if your protocol wait for ACK from the receiving end you should take that into account.
you should also consider time necessary to read/build/retrieve data to send and really send it over the wire. Depending of what you are doing it can be negligible or take several seconds...
If you give some more details it will be easier to diagnostic the source of the problem. sleep as you believe (it is indeed a really poor timer) or some other part of your system.
If your dump is large, I will bet that the additional time comes from reading data and sending it over the wire. You should mesure time consumed in the sending process (reading time before and after finishing sending).
If this is indeed the source of the additional time, you just have to remove that time from the next time to wait.
Example: Sending the previous block of data took 4s, the next block is 10s later, but as you allready consumed 4s, you just wait for 6s.
sleep is still a quite imprecise timer and obviously the above mechanism won't work if sending time is larger than delay between sendings, but you get the idea.
Correction sleep is not so bad in windows environment as it is in unixes. Accuracy of windows sleep is millisecond, accuracy of unix sleep is second. If you do not need high precision timing (and if network is involved high precision timing is out of reach anyway) sleep should be ok.
Any modern multitask OS's scheduler will not guarantee any exact timings to any user apps.
You can try to assign 'realtime' priority to your app some way, from a windows task manager for instance. And see if it helps.
Another solution is to implement a 'controlled' sleep, i.e. sleep a series of 500ms, checking current timestamp between them. so, if your all will sleep a 1s instead of 500ms at some step - you will notice it and not do additional sleep(500ms).
Try out a Multimedia Timer. It is about as accurate as you can get on a Windows system. There is a good article on CodeProject about them.
Sleep function can take longer than requested, but never less. Use winapi timer functions to get one function called-back in a interval from now.
You could also use the windows task scheduler, but that's going outside programmatic standalone options.